Download Predicting phenology by integrating ecology, evolution and climate

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Scientific opinion on climate change wikipedia , lookup

Climate change and poverty wikipedia , lookup

Effects of global warming on humans wikipedia , lookup

Public opinion on global warming wikipedia , lookup

IPCC Fourth Assessment Report wikipedia , lookup

Surveys of scientists' views on climate change wikipedia , lookup

Years of Living Dangerously wikipedia , lookup

Climate change, industry and society wikipedia , lookup

Effects of global warming on Australia wikipedia , lookup

Transcript
Global Change Biology (2011) 17, 3633–3643, doi: 10.1111/j.1365-2486.2011.02515.x
REVIEW
Predicting phenology by integrating ecology, evolution
and climate science
STEPHANIE PAU*, ELIZABETH M. WOLKOVICH†, BENJAMIN I. COOK‡, T. JONATHAN
D A V I E S § , N A T H A N J . B . K R A F T ¶ , K J E L L B O L M G R E N * * , J U L I O L . B E T A N C O U R T † † and
ELSA E. CLELAND†
*National Center for Ecological Analysis & Synthesis, 735 State Street, Santa Barbara, CA 93101, USA, †Division of Biological
Sciences, University of California – San Diego, 9500 Gilman Drive #0116, La Jolla, CA 92130, USA, ‡NASA Goddard Institute
for Space Studies, 2880 Broadway, New York, NY 10025, USA, §Department of Biology, McGill University, 1205 ave Docteur
Penfield, Montreal, Quebec H3A 1B1, Canada, ¶Biodiversity Research Centre, University of British Columbia, 6270 University
Blvd, Vancouver, BC V6T 1Z4, Canada, **Department of Biology, Lund University, Ecology Bldg, SE-223 62, Lund, Sweden,
††U.S. Geological Survey, National Research Program, 1955 E. 6th Street, Tucson, AZ 85719, USA
Abstract
Forecasting how species and ecosystems will respond to climate change has been a major aim of ecology in recent
years. Much of this research has focused on phenology – the timing of life-history events. Phenology has well-demonstrated links to climate, from genetic to landscape scales; yet our ability to explain and predict variation in phenology
across species, habitats and time remains poor. Here, we outline how merging approaches from ecology, climate science and evolutionary biology can advance research on phenological responses to climate variability. Using insight
into seasonal and interannual climate variability combined with niche theory and community phylogenetics, we
develop a predictive approach for species’ reponses to changing climate. Our approach predicts that species occupying higher latitudes or the early growing season should be most sensitive to climate and have the most phylogenetically conserved phenologies. We further predict that temperate species will respond to climate change by shifting in
time, while tropical species will respond by shifting space, or by evolving. Although we focus here on plant phenology, our approach is broadly applicable to ecological research of plant responses to climate variability.
Keywords: environmental filtering, growing-degree day models, niche conservatism, photoperiod, temperature sensitivity, temporal niche
Received 29 March 2011; revised version received 11 July 2011 and accepted 21 July 2011
Introduction
Accurate forecasting of how species will respond to climate change requires perspectives fr‘om the fields of
ecology, climatology and evolutionary biology (Jackson
et al., 2009). Synthesizing these perspectives, however,
requires reconciling both fundamental differences in
the temporal and spatial scales at which ecological and
evolutionary processes can operate, as well as divergent views of the principal drivers underlying
responses (Benton, 2009). This tension between research
fields is particularly evident in the study of plant phenology, defined as the timing of periodic life-history
events such as leaf budburst or first flower.
S. Pau and E.M. Wolkovich Authors contributed equally to the work.
Correspondence: S. Pau, tel. + 805 892 2517, fax + 805 892 2510,
e-mail: [email protected]
© 2011 Blackwell Publishing Ltd
Phenology is strongly linked to climate – which, for
the purpose of this article, we define as the composite
of generally prevailing weather conditions (e.g. temperature and precipitation) at a site or over a region, for
some defined period of time (e.g. months, seasons and
years). The magnitude and direction of plant species’
phenological responses to climate cues have widespread consequences for trophic interactions, ecosystem services, and our ability to predict the shape of
future communities, which feed back into important
biosphere–atmosphere interactions (Cleland et al., 2007;
Parmesan, 2007). Accurately forecasting phenology is
thus a current objective in many fields, but these fields
have widely divergent perspectives. Community ecologists have focused on localized, short-term (generally
1–3 years) studies that emphasize pairwise species
interactions, trophic mismatches or competition for
resources (Sargent & Ackerly, 2008; Miller-Rushing
et al., 2010; Thackeray et al., 2010). Climate-focused
3633
3634 S . P A U et al.
studies rely on long-term (i.e. decades) and synopticscale (ecosystems to biomes) observations to identify
shared climatic signals in phenology (Menzel et al.,
2006; Schwartz et al., 2006), or they use phenology as a
constraint on biogeochemical feedbacks between the
biosphere and atmosphere (Peñuelas et al., 2009; Richardson et al., 2009). Species-specific phenological studies have been scaled-up in the context of spatial and
temporal shifts in range sizes through process-based
models (Chuine & Beaubien, 2001; Morin et al., 2009),
but these models still lack a community context.
Demonstrating how community-level processes influence the larger-scale functional roles of ecosystems
remains a challenge. Trait-based approaches that consider how the traits of organisms turn over along abiotic
gradients provide a bridge between processes at the
community level and global change predictions (McGill
et al., 2006; Ackerly & Cornwell, 2007; Suding et al.,
2008). Phenological traits, such as flowering time and
plant sensitivities to climate (e.g. degree of shift in phenology with shift in temperature), can be mapped onto
phylogenetic trees. Phylogenetic methods provide an
integrated approach to predicting the phenology of
many species, and will allow us to address critical questions on how constrained or variable phenology may be
across species, sites, and time in response to climate
variability (Ollerton & Lack, 1992; Willis et al., 2008).
Here we briefly review how climatology and ecology
have traditionally approached phenological research,
and in particular where these fields diverge. We consider how different scales and paradigms have shaped
findings, including which environmental cues govern
phenology – especially the widely studied events of leafburst and flowering. We propose that a more integrative
approach drawing on niche theory and community phylogenetics could use site- and species-specific plant
responses to make broad-scale predictions for many
species. We further emphasize how shifting abiotic and
biotic forces, which vary with seasonal and interannual
climate variability, should shape phenologies across
space and time. Our framework provides predictions
for current patterns of plant sensitivities to climate, has
direct ties to how species, communities, and ecosystems
will respond to future climate change, and should be
testable with current climate and phenology data.
Environmental cues: linking climate to phenology
across scales
For phenology, as with many fields of ecology today,
the holy grail is to unify our understanding of variation
across scales, linking genetic studies to the expansive
spatial and temporal scales of natural systems (Fig. 1).
At any scale, a fundamental challenge is identifying the
Fig. 1 Predicting how species respond to climate requires a
merging of perspectives and matching climate data to relevant
biological scales. While genetic-physiological studies have provided detailed information for several species on daily to seasonal scales, they are not easily applied to other species.
Process-based and integrative climate models have linked field
phenology to daily, seasonal and large-scale climate metrics.
There is a wide range of ecological-statistical models, however
studies on larger spatial scales tend to ignore community-level
variation where phylogenetic studies may offer insight into
community structure. Almost no efforts have examined how
plant phenology has responded to previous long-term shifts in
environmental cues. (Shading is meant to delineate overlapping
boxes representing model types and the scales they typically
address).
suite of environmental cues that initiate biological processes. But different fields of study have divergent findings regarding the relative importance among
phenological cues such as irradiance, temperature or
precipitation. This disparity is potentially due to inherent differences in the way that researchers from different fields approach the study of phenology. For
© 2011 Blackwell Publishing Ltd, Global Change Biology, 17, 3633–3643
P R E D I C T I N G P H E N O L O G Y 3635
example, we reviewed studies that sought to identify
environmental cues for flowering or budburst and
found that at the genetic and physiological levels,
approximately half of all studies (51%, see Appendix S1) identify photoperiod or irradiance cues, with
temperature following closely behind (32%). In contrast, ecological (field or plot-scale) and climatological
studies overwhelmingly find temperature cues across
species and latitudes (86%) while <3% cite photoperiod
or irradiance. All fields, however, find approximately
the same proportion of cues due to precipitation, at
about 10%. Resolving these differences is key to predicting biological responses to climate change because
accurately forecasting phenology depends on identifying the correct cues.
Climatologists (or more specifically, biometereologists or bioclimatologists) have viewed phenology largely as an adaptation to avoid environmental stress,
especially cold and drought. By assuming that the phenological response is broadly synchronous and correlated to climate, climatologists have treated plant
species effectively as permuted meteorological stations,
where one or two event dates (e.g. leafing and flowering) represent an integrated metric of climate over the
preceding days and seasons. With this perspective
climate-focused studies have been able to work on large
spatial scales, matching regional and continental-scale
climate patterns to phenology. In particular, ‘growingdegree day’ models have proved highly accurate at predicting start-of-season (‘spring’) phenology. These statistical models typically use daily temperature data to
capture chilling requirements and heat accumulations
over the course of a season (Schwartz et al., 2006). However, such models may not detect when multiple environmental cues are required to initiate a phenological
event, nor detect cues that are relatively static across
space, such as photoperiod. Process-based models
(Chuine & Beaubien, 2001) attempt to work around
these issues by combining species or population-level
information on thresholds and dormancy, usually
derived from growth-chamber experiments, using
time-series approaches. These more complex models
have provided insight into how phenology may determine ranges of several northern temperate tree species
(Morin et al., 2008) and thus offer promise for scaling
from physiological processes to patterns of species
composition over the landscape. In addition, climate
indices such as the NAO (North Atlantic Oscillation)
Fig. 2 Growing season length scales inversely with latitude. Thus, compared to northern areas, lower latitudes provide a longer period
of time each year for species to be active (i.e. greater temporal niche axis, curves represent idealized species niches for simple communities). Alongside this, we predict that the relative importance of biotic vs. abiotic drivers varies with latitude such that aseasonal tropical
communities have mainly biotic drivers of phenology because mis-timing would likely lead to mis-match with pollinators or competitors, not highly unfavorable abiotic conditions (but see Prediction 3 and Fig. 5). In contrast, the phenology of arctic communities is driven mainly by abiotic forces, and species should be highly sensitive to climate to avoid heavy abiotic costs associated with mis-timed
phenological events. Additionally, for seasonal systems the relative role of abiotic and biotic forces would vary such that abiotic selective forces predominate for species active at the start or end of growing seasons, when the costs of small mis-calculations should promote high sensitivity to climate. In contrast, species that leaf or bloom during periods of lower abiotic stress may have less sensitive
climate cues but track strongly to other cues which allow them to map onto periods of low competition for soil or pollinator resources.
Carefully testing these predictions, however, requires far more long-term tropical data than currently available, because phenological
records are generally focused on temperate latitudes (black points, references given in Appendix S2).
© 2011 Blackwell Publishing Ltd, Global Change Biology, 17, 3633–3643
3636 S . P A U et al.
may have greater power to predict phenology than single climate variables (e.g. temperature and precipitation) because they integrate many weather variables
that influence phenology over seasons and interannual
timescales (Beaubien & Freeland, 2000; Stenseth & Mysterud, 2005; Vicente-Serrano et al., 2006) and can capture abrupt shifts in climate (Cayan et al. 2001).
A major limitation of these climatological modeling
approaches is their reliance on an underlying assumption of both temporal and spatial stationarity – i.e. phenological responses to climate that are stable and
consistent through time and across space. Phenological
models have largely relied on data from temperate
mid-latitudes (Fig. 2) and on analyses from networks of
cultivated and clonal species (Schwartz & Reiter, 2000;
Menzel et al., 2006). However, phenology–climate relationships derived from these clonal studies may not
scale easily to natural populations where species’ sensitivities to climate can vary across individuals, sites,
communities, and climate gradients (Olsson & Agren,
2002; Jentsch et al., 2009). Assessing the validity of the
stationarity assumption will be critical for models to
accurately predict ecological responses to climate
change globally.
Climatologists have generally de-emphasized intraand inter-specific variation in phenology, in part due to
the lack of long-term, spatially distributed data for all
but a few species. Such de-emphasis has advantages,
allowing climate-focused studies to infer patterns
across broad spatial and temporal scales, but with the
implicit assumption that variation among and within
species do not scale up in a meaningful way to affect
ecosystem functions that influence climate processes.
Where broad-scale climatological studies may fall short
– by ignoring inter- and intra-specific differences –
physiological research into phenology has excelled for
a handful of model species. While a number of crop
and lab model species have been studied (see Appendix S2), such as Arabidopsis thalina (Mouradov et al.,
2002), enormous efforts are required to produce a
model for a single species (Wilczek et al., 2009). Scaling
up from lab and crop species to natural systems is currently impossible. Thus even with carefully mapped
physiological pathways, a community context is still
crucial to predicting phenology in natural communities.
Ecological research emphasizes how interspecific
interactions and variation can drive phenology. In particular, phenology provides a mechanism by which cooccuring species can reduce competition by partitioning resources through time – the temporal niche (Gotelli & Graves, 1996) – hence the community context of
phenology becomes important. This prediction is supported by work showing that a species’ phenology may
be sensitive to local community composition (Lack,
1982). Ecologists have typically regarded climate as a
source of cues underlying phenological events upon
which selection acts, for example, to maximize germination rates, match pollinators, or reduce resource competition (van Schaik et al., 1993). Because selection
might act independently upon multiple life-history
traits, phenology may be correlated with, or constrained by, a number of reproductive characters,
including pollination mode (Rabinowitz et al., 1981),
seed set and seed dispersal (Mazer, 1990; Oberrath &
Böhning-Gaese, 2002; Bolmgren & Cowan, 2008).
In order to make significant progress from documenting change to predicting and forecasting phenology,
researchers must draw on the strengths of ecological
and climatological perspectives, while embracing new
approaches that incorporate evolutionary change.
Selection and adaptation are critical to understanding
differential and non-stationary species responses to climate change (Hoffman & Sgrò, 2011), but a framework
for mapping phenological variation across species and
sites is first needed.
A multi-species approach: integrating phylogeny
and traits into phenological research
Evolutionary trees and mapping phenological traits
An evolutionary perspective provides a multi-species
approach for plant phenology research. Because
physiological pathways and responses are likely to
be evolutionary conserved (i.e. close relatives share
similar traits), it is possible to generate predictive
models using information on the evolutionary relationships among species – their phylogeny. Phylogeny may be an especially powerful approach for
understanding phenology because it provides a simple method to integrate species differences across
multiple traits, which may have complex underlying
physiological pathways. Phylogenetic approaches can
map both raw observational phenology data, such as
flowering dates, as well as model estimates of how
strongly species cue to different environmental variables such as temperature (Bolmgren & Cowan, 2008;
Willis et al., 2008; Davis et al., 2010). In addition,
because phylogeny allows us to infer the evolutionary dynamics of trait changes it can also help
address questions regarding the underlying physiological pathways that determine phenology, as well
as the ecological and evolutionary pressures that
structure communities and drive variation in the timing of events. Importantly, by placing phenology
within an historical evolutionary framework, we can
also project forward to predict adaptive potential in
© 2011 Blackwell Publishing Ltd, Global Change Biology, 17, 3633–3643
P R E D I C T I N G P H E N O L O G Y 3637
response to future climate change (Hoffman & Sgrò,
2011).
The accuracy with which phylogeny predicts phenology can be assessed by evaluating the strength of evolutionary niche conservatism in phenological responses
(as a proxy for the conservatism of the underlying traits
that determine responses). While evolutionary niche
conservatism is increasingly recognized as a pervasive
phenomenon with broad ecological implications (Wiens
et al., 2010), there has been some debate as to its definition (Losos, 2008). For purposes here, we partition
niche conservatism into two components: first, the
strength of the covariance between the evolutionary
distances between taxa and their difference in trait values, referred to as phylogenetic signal (Blomberg et al.,
2003). Second, an evolutionary rate component, describing the velocity of change along the branches of the
evolutionary tree (Ackerly, 2009). Importantly, strength
of phylogenetic signal for unconstrained traits (e.g.
traits that have not yet approached the bounds of evolutionary limits) is independent from evolutionary rates
(Revell et al., 2008). Therefore phenology might map
closely onto phylogeny (strong phylogenetic signal), so
that evolutionary divergence is a good predictor of difference in phenology, but species differences may be
large or small depending upon the evolutionary rate of
change. Accurately tracing the evolution of phenology
requires information on both signal and rates – to date,
we have little information on either (but see Willis et al.,
2008).
Phylogenetic approaches also provide a means to test
for temporal niche differentiation in communities. For
example, if closely related species share similar phenological traits, then at local spatial scales species may
partition themselves through time to reduce competition, producing a phylogenetically over-dispersed pattern (species less related than expected) (Webb, 2000;
Cavender-Bares et al., 2004). In contrast, at larger spatial scales, environmental filtering, defined as abiotic
limitations on growth and/or establishment (van der
Valk, 1981; Cavender-Bares et al. 2009), may produce
communities of species that share similar traits related
to environmental tolerance (e.g. traits that optimize
phenology to the growing season) and therefore would
produce under-dispersed assemblages (Webb et al.,
2002; Kraft et al., 2007). Importantly, predictions are
senstive to the mode of trait evolution, for example,
phylogenetic community structure might be absent if
phylogentic signal in phenology is weak or if phenological traits are evolutionarily convergent (Webb et al.,
2002). Therefore, it is critical to construct robust phylogenetic hypotheses of trait change.
While integrating phylogeny into phenological
research provides important information on constraints
© 2011 Blackwell Publishing Ltd, Global Change Biology, 17, 3633–3643
and flexibility across species, studies using contemporary methods are uncommon and have only been conducted at the single community or sub-community
level (Fig. 1). However such studies, might prove particularly powerful for detecting evolutionary trends
because members of a community share a common
environment, and hence experience the same suite of
environmental cues. Currently, weight-of-evidence suggests that flowering time shows strong phylogenetic
signal and is highly conserved within species. In the
largest analysis to date, including data on >5000 species, Kochmer & Handel (1986) found strong congruence in flowering times for species shared between the
floras of North Carolina and Japan. Evidence for evolutionary niche conservatism has been found in several
species-rich floras, including the Cape of South Africa
(Johnson, 1993) and some tropical forests (Wright &
Calderon, 1995). In a recent study of one northern US
plant community, Willis et al. (2008) found species’
phenological sensitivities to spring temperatures and
extinction risk were evolutionarily conserved. If such
patterns generalize widely, phylogeny will be a useful
tool for predicting species sensitivities in less well-studied communities. However, using a long-term record of
UK flora, Davis et al. (2010) found evidence for phylogenetic signal in climate tracking only when some
clades were excluded, indicating community context
might be important. Further research across diverse
systems is therefore vital.
Predictions across space and time
Synthesizing perspectives from ecology and evolutionary biology with climate science should advance
research towards developing a framework for predicting phenological responses to climate. The predictive
framework we present here is based on evolutionary
relationships between coexisting species and exploring
the temporal niche as a reflection of latitudinal variation in the growing season, the seasonal transition
length and interannual variability. Our predictions of
plant phenology consider fitness costs to mis-timing,
and specifically we suggest that the relative importance
of abiotic forces, such as environmental filtering, and
biotic forces, especially niche-based processes (e.g.
competition), will variably shape species and community phenologies across growing seasons and latitudes
(Fig. 2).
Major drivers of plant phenology: abiotic vs. biotic
drivers
Species are under continuous selective pressure to
match their phenologies to favorable environmental
3638 S . P A U et al.
conditions and positive biotic interactions to reduce fitness costs associated with mis-timing. Costs of mistimed phenology associated with abiotic drivers should
be high (Schwartz et al., 2006): blooming too early or
late can lead to death or extensive tissue loss due to
frost in temperate climates (Inouye, 2008) or drought in
semi-arid systems. In contrast, costs associated with
biotic drivers may be lower over short-time scales
(Ollerton & Lack, 1992). For example, if a species flowers during the growing season but at the wrong period
it may face increased competition for soil resources or
reduced pollination, which may reduce growth and
reproductive output. The relative strength and proportion of species within a community governed by these
drivers should vary according to local environmental
conditions and species’ ecologies: in particular we suggest three metrics of climate variability may be important.
How growing season, seasonal transition length, and
interannual variability may shape the phenologies of
species and communities
Prediction 1: Early season species’ phenologies should be
more sensitive to abiotic forces, whereas mid-growing season
species’ phenologies should be governed by biotic forces. As
growing season length increases towards the tropics the relative within-season weight of abiotic vs. biotic forces on plant
phenology should decrease.
The growing season defines the window of time in
which plant growth is possible, i.e. the fundamental
temporal niche. While some tropical wet forests have
year-round seasons (i.e. they are aseasonal), with argu-
ably many temporal niches for species to occupy (Mittelbach et al., 2007), most habitats, including many
tropical forests, are seasonal. Therefore most species
must time their growth carefully to capture optimal conditions of sufficient warmth, irradiance and soil moisture (Larcher, 2003), but also to minimize competition
for limited resourses. In seasonal environments, phenology determines a species’ ability to establish and persist
within the local temporal niche (i.e. environmental filtering in relation to phenology, Fig. 3a). The phenology
of a species relative to other members of the community
should have important ramifications for species interactions, such as competition for resources (favoring
temporal displacement) or facilitation via shared pollinators (favoring temporal convergence) (Fig. 3b).
The period within the growing season that a species
occupies should additionally determine the relative
strength of abiotic vs. biotic forces driving phenology
and thus the types of cues used to time events (Fig. 2).
The strength of abiotic forces often varies within a
growing season: for example, there should be strong
abiotic forces at the beginning of the growing season in
temperate environments and the mid-growing season
for environments characterized by mid-summer
droughts because mis-timing has large fitness costs. For
species with temporal niches occupying these portions
of the growing season, the heavy costs of small mis-calculations should promote high sensitivity to climate
(producing a flexible phenology) because the calendar
day of these abiotic forces varies dynamically in most
systems from year to year.
In contrast, species occupying other portions of the
growing season may be less sensitive to climate and
Fig. 3 Phylogenetic patterns within communities may reflect both broad scale environmental filtering (a) as well as local-scale partitioning of time in response to community context (b). For example, in temperate systems species filling the same general ecological
niche (or guild) may sort into environments using common chilling requirements (e.g. a habitat filter, a), resulting in a phylogenetically
clustered pattern if chilling requirements are phylogenetically conserved. However, at smaller spatial scales within a community (b),
species may subdivide temporal niche space using more subtle environmental cues, such as varying temperature or precipitation triggers.
© 2011 Blackwell Publishing Ltd, Global Change Biology, 17, 3633–3643
P R E D I C T I N G P H E N O L O G Y 3639
Fig. 4 The speed of transition into the growing season (x-axis)
may fundamentally affect the phenologies of species within biomes. Species, especially those that are active during rapid transitions, should be most plastic with climate in areas with the
shortest seasons (represented by darker shading, e.g. tundra).
could use more static cues such as photoperiod. For
example, in many mesic temperate systems, early season species should attune to abiotic cues so as to prevent growth during the winter or very early spring.
However, mid-growing season species should be governed by biotic forces, such as competition for resources
(Morales et al., 2005), because potential for species phenological overlap is highest in the mid-season (i.e. middomain effect) and the risk of encountering adverse
growing conditions is lowest. Thus, selection pressure
for climate sensitivity may be low and static cues such
as photoperiod might dominate (Calle et al., 2010),
allowing for consistent partitioning of the temporal
niche from year-to-year. Evidence that early season species are most sensitive to climate supports this hypothesis (Menzel et al., 2006), but there is little work that we
are aware of testing how cues vary along a growing
season (Ollerton & Lack, 1992) and no general predictions for which cues species may use to time their
growth to reduce competition (further discussion
below). We predict that as growing seasons lengthen
towards the tropics the relative within-season weight of
abiotic vs. biotic forces on plant phenology should
decrease, such that in high latitude systems almost all
species must be sensitive to climate to capture the
shorter growing season, while in tropical systems the
majority of species may use static cues (Fig. 2).
© 2011 Blackwell Publishing Ltd, Global Change Biology, 17, 3633–3643
Prediction 2: Species phenologies occupying relatively long
gradual seasonal transition periods should be sensitive to
climate. Species timed to rapid transitions may be less sensitive to climate because of high competition.
For many environments the temporal niche space
may additionally be defined by the seasonal transition
length (the absolute time it takes for a system to transition into and out of its growing season). In most environments, the majority of the growing season is defined
by a period of relative climatic stability when most species grow and flower. Some systems have extremely
rapid, consistent transition periods such as tropical
monsoon forests (Elliott et al., 2006; Williams et al.,
2008), while many habitats have longer transitions,
such as in most temperate biomes (Fig. 4). Fewer species often grow and flower during the transitional period in seasonal environments as compared to the midseason (van Schaik et al., 1993; Morales et al., 2005),
thus we predict that transitional species face lower
competition for resources. Yet, these species should
experience high costs if they mis-time their growth (in
conditions too cold or dry for tissue growth). Thus species occupying gradual seasonal transition periods
should cue to climate and flexibly shift their phenologies between years, even within biomes where the total
growing season is long. In contrast, wet tropical biomes, especially monsoon systems, often have very short
transitions, during which many species begin their
growth and reproduction rapidly. In such systems
cuing to photoperiod or other non-climate cues may be
more advantageous to avoid competition (Elliott et al.,
2006) and could result in low variation in early season
timing. Across biomes, the intersection of growing season length (see Prediction 1) and the seasonal transition
length may shape how plants respond to climate variation (Fig. 4).
Prediction 3: Species phenologies in environments with low
interannual variability should respond more strongly to biotic forces, whereas species in environments experiencing high
interannual variability should be more sensitive to abiotic
forces.
Interannual variability in temperature or moisture
should also affect the relative importance of abiotic vs.
biotic drivers by influencing the optimal bet-hedging
strategy, a strategy that describes plant responses in the
face of environmental uncertainty (Fig. 5). In areas with
long growing seasons and relatively low interannual
variability, such as many low latitude regions (Fig. 6),
costs associated with mis-timing are low and biotic
forces are therefore expected to drive phenology. In
systems with high interannual variability however,
3640 S . P A U et al.
Fig. 5 Interannual variability in the mean temperature (a, standard deviation in °C) and precipitation (b, coefficient of variation – unitless) combined with seasonality of temperature (c, in °C) and precipitation (d, unitless) varies across the globe and should predict the
strength of different phenological cues. Plant species in tropical regions of the Americas and some parts of Asia and Africa experience
relatively low interannual variability and seasonality and thus may use distinctly different cues than species in regions of high seasonality and variability in rainfall (Mediterranean regions, parts of Asia and Australia). Across temperate and arctic regions – where the
majority of phenology work has been conducted – high seasonality and variability in temperature dominates. This makes extrapolating
from these regions to other areas of the globe, which have distinctly different patterns of interannual variability and seasonality, difficult. Note that seasonality scales inversely with growing season length, such that aseasonal systems have year-round growing seasons
and highly seasonal habitats (in temperature or precipitation) have comparatively short seasons for plant growth and reproduction.
Data sources and complete methods are given in Appendix S2.
Fig. 6 Growing season length and interannual variability in climate may predict the most common phenological strategy for most species in a habitat. Abiotic forces (light shading) should restrict membership in a community to species with minimal cues, and thus
where stabilizing selection should be strong. In contrast other environments may be dominantly governed by biotic forces (dark shading) and thus there would be comparatively weak selection on phenological cues.
climate sensitivity may be more important. Temperature and precipitation cues may then interact with or
mask photoperiod cues to help species optimally time
growth and reproduction. In such systems multiple
strategies that vary with the frequency of life-history
events may occur: species that bloom often (e.g. annually) would place consistent but smaller bets, while
other species may bloom supra-annually, avoiding
years of sub-optimal climate and expending high
resources in optimal years (Venable, 2007). Such strategies appear to underlie many semi-arid systems (Venable, 2007) and monsoon tropical forests (Sakai, 2001),
and should be further dependent on the generation
time of species (Chesson, 2000). In addition, species
placing larger bets to cue to optimal conditions that
occur less frequently may rely upon multiple cues,
© 2011 Blackwell Publishing Ltd, Global Change Biology, 17, 3633–3643
P R E D I C T I N G P H E N O L O G Y 3641
whereas species placing smaller bets might converge
on single cues that are only good predictors on average.
Variation in evolutionary rates in phenological traits
Prediction 4: Phenological traits responding to cues associated with temporal niche-partitioning, such as photoperiod,
should exhibit faster evolutionary rates, whereas responses to
temperature and precipitation should exhibit slower evolutionary rates.
Our framework, based on the three climate metrics
described above, assumes variation in the costs of mistimed phenology associated with different cues. A difference in costs associated with mis-timed phenology
might translate into variation in evolutionary rates –
the rate at which species’ phenotypes diverge over
time. Specifically, if photoperiod is associated with temporal niche partitioning to avoid competition for
resources, the costs of mis-timing may be relatively
low. We therefore predict greater evolutionary lability
in plant responses to photoperiod cues, which may
translate into faster evolutionary rates. In contrast,
responses to temperature or precipitation cues might
evolve slowly (Fig. 6) because they are associated with
abiotic drivers of selection and thus likely under strong
stabilizing selection (Levin, 2006). Where strategies are
mixed, we predict responses to cues might also be
mixed. For example, while we predict some lability
around the timing of annual blooming species,
responses to cues for supra-annual species should
evolve slowly since blooming in sub-optimal years
would impose high costs. To date there have been
almost no studies exploring variation in evolutionary
rates of traits associated with response to phenological
cues (see Martin et al., 2009). Because responses should
be apparent directly as variation in phenology, it
should be possible to evaluate predictions with current
data given reasonable estimates of phylogeny.
Predicting the cues and their conservatism across
species and biomes has clear relevance to forecasting
plant responses to climate change (Fig. 6). Species that
track climate variables closely should be most able to
adjust their phenologies with climate change and face
minimal population changes (Visser, 2008). This suggests species in high latitudes (Fig. 5) should cope well
with climate change. However, because we predict that
such plasticity is under strong stabilizing selection
there may be little genotypic variation underlying this
plasticity. Thus, if climate shifts beyond conditions for
which species are adapted and cues or cue sequences
break down, rapid evolutionary shifts might not be
possible. In contrast, species that occupy less variable
habitats such as the aseasonal tropics may have little
ability to track climate, and thus should instead shift
© 2011 Blackwell Publishing Ltd, Global Change Biology, 17, 3633–3643
their ranges or face population declines. However, if
biotic forces structure species’ phenologies in such
communities (see Predictions 1–3), then they may
also have high variation in phenology, suggesting
that some species may be optimally timed for new
climate regimes, but we risk losing species at both
the trailing and leading edges of the climate window.
Cues underlying tropical phenology have allowed
most species in the past to remain extant by shifting
their ranges in response to global shifts in photoperiod and climate associated with variations in the
earth’s orbit and tilt (Jansson & Dynesius, 2002).
Research is needed, however, to test how species will
respond to rapidly changing cues independent from
photoperiod (Fig. 1), and to assess the risk that climate change might disrupt cascading cues (both biotic and abiotic).
Current opportunities in testing predictions
Testing these predictions certainly requires more field
and lab-based studies. However, we can make large
advances using current long-term records of species phenology by integrating frameworks from climatology and
ecology. Ecological predictions may improve by matching organismal scales with climate variables (Fig. 1).
While ecologists have used a wide variety of different
climate variables and models (Diekmann, 1996; Post &
Stenseth, 1999), recent work has often used monthly and
annual means, which operate on human calendar scales,
instead of on biological scales (Yang & Rudolf, 2010).
With the increased availability of daily climate data (Peterson & Vose, 1997) future ecological research should
use daily models that have clear ties to plants’ circadian
timescales (Dodd et al., 2005) and to how climate change
has influenced daily minima and maxima disproportionately (Vose et al., 2005). In turn, climate science could
benefit from greater focus on intra- and inter-specific
variation, resulting in more accurate predictions of biological responses to climate change. For example
Dynamic Global Vegetation Models as components of
coupled climate-carbon models are being evolved to
realistically represent biodiversity and competition
among individual forest species, rather than assume
some average forest phenology (Lichstein et al., 2010).
Clearly, more research is needed across tropical and
semi-arid systems (Fig. 2), and across time at the middle and end of growing seasons. In addition, while we
have focused here on photoperiod, precipitation and
temperature, a number of other cues may modulate
phenology, and these might be more important in nontemperate systems. For example, in some parts of the
tropics phenology may be most sensitive to variation in
solar insolation produced by changes in cloud cover
3642 S . P A U et al.
and drought (Huete et al., 2006; Asner & Alencar, 2010),
or tree water potential, which is not always easily
related to recent precipitation (Reich & Borchert, 1984).
Our predictions also point toward the need for a far better understanding of which cues species use to partition
themselves in temporal niches within a community. We
have assumed that photoperiod may be important to
temporal niche partitioning, because it should allow
species to consistently partition the growing season
between years, and because evidence suggests it is a
dominant cue for species with lower abiotic stress (Calle et al., 2010). Other research however, suggests species
may use a variety of mechanisms (Borchert et al., 2004;
Calle et al., 2010). Additionally, environmental cues
underlying end of growing season phenology (e.g. leaf
coloring, leaf abscission, etc.) are poorly understood
(Menzel et al., 2006), but appear to be less variable from
year to year and less sensitive to temperature cues,
especially when compared to start of growing season
metrics (Barr et al., 2004). Our poor understanding is
especially disconcerting, given the importance of end of
growing season timing for ecosystem functioning,
including net ecosystem productivity (Angert et al.,
2005). Thus, even within the highly sampled temperate
midlatitudes (Fig. 2), there are still large gaps in our
understanding of phenology-climate connections.
Acknowledgements
This work was conducted as part of the Forecasting Phenology
working group supported by the National Center for Ecological
Analysis and Synthesis (NCEAS), a Center funded by NSF
(Grant #EF-0553768), the University of California, Santa Barbara,
and the State of California, and was supported by the USA
National Phenology Network and its NSF RCN grant (IOS0639794) and conducted while EMW was an NSF Postdoctoral
Research Fellow in Biology (DBI-0905806), SP was an NCEAS
postdoctoral associate and while NJBK was supported by the
NSERC CREATE training program in biodiversity research.
Please see Appendix S3 for author contributions. We thank A.
Rogstad and J. Weltzin for assistance in reviewing cues across
fields, H. Kharouba, S. Mazer, C. Parmesan, K. Gerst, and J.
Morisette for comments on an earlier draft, and thank all members of the working group for discussion and perspectives
across fields.
References
Ackerly DD (2009) Colloquium Papers: Conservatism and diversification of plant
functional traits: evolutionary rates versus phylogenetic signal. Proceedings of the
National Academy of Sciences, 106, 19699–19706.
Ackerly DD, Cornwell WK (2007) A trait-based approach to community assembly:
partitioning of species trait values into within- and among-community components. Ecology Letters, 10, 135–145.
Angert A, Biraud S, Bonfils C et al. (2005) Drier summers cancel out the CO2 uptake
enhancement induced by warmer springs. Proceedings of the National Academy of
Sciences USA, 102, 10823–10827.
Asner GP, Alencar A (2010) Drought impacts on the Amazon forest: the remote sensing perspective. New Phytologist, 187, 569–578.
Barr AG, Black T, Hogg E, Kljun N, Morgenstern K, Nesic Z (2004) Inter-annual variability in the leaf area index of a boreal aspen-hazelnut forest in relation to net ecosystem production. Agricultural and Forest Meteorology, 126, 237–255.
Beaubien EG, Freeland HJ (2000) Spring phenology trends in Alberta, Canada: links
to ocean temperature. International Journal of Biometeorology, 44, 53–59.
Benton MJ (2009) The Red Queen and the Court Jester: species diversity and the role
of biotic and abiotic factors through time. Science, 323, 728–732.
Blomberg SP, Garland T, Ives AR (2003) Testing for phylogenetic signal in comparative data: behavioral traits are more labile. Evolution, 57, 717–745.
Bolmgren K, Cowan PD (2008) Time-size tradeoffs: a phylogenetic comparative study
of flowering time, plant height and seed mass in a north-temperate flora. Oikos,
117, 424–429.
Borchert R, Meyer SA, Felger RS, Porter-Bolland L (2004) Environmental control of
flowering periodicity in Costa Rican and Mexican tropical dry forests. Global Ecology and Biogeography, 13, 409–425.
Calle Z, Schlumpberger BO, Piedrahita L, Leftin A, Hammer SA, Tye A et al. (2010)
Seasonal variation in daily insolation induces synchronous bud break and flowering in the tropics. Trees, 24, 865–877.
Cavender-Bares J, Ackerly DD, Baum DA, Bazzaz FA (2004) Phylogenetic overdispersion in Floridian oak communities. American Naturalist, 163, 823–843.
Cavender-Bares J, Kozak KH, Fine PVA, Kembel SW (2009) The merging of community ecology and phylogenetic biology. Ecology Letters, 12, 693–715.
Cayan DR, Kammerdiener SA, Dettinger MD, Caprio JM, Peterson DH (2001)
Changes in the onset of spring in the western United States. Bulletin of the American
Meteorological Society, 82, 399–415.
Chesson P (2000) Mechanisms of maintenance of species diversity. Annual Review of
Ecology and Systematics, 31, 343–366.
Chuine I, Beaubien EG (2001) Phenology is a major determinant of tree species range.
Ecology Letters, 4, 500–510.
Cleland EE, Chuine I, Menzel A, Mooney HA, Schwartz MD (2007) Shifting plant
phenology in response to global change. Trends in Ecology & Evolution, 22, 357–365.
Davis CC, Willis CG, Primack RB, Miller-Rushing AJ (2010) The importance of phylogeny to the study of phenological response to global climate change. Philosophical
Transactions of the Royal Society B: Biological Sciences, 365, 3201–3213.
Diekmann M (1996) Relationship between flowering phenology of perennial herbs
and meteorological data in deciduous forests of Sweden. Botany, 74, 528–537.
Dodd AN, Salathia N, Hall A, Kévei E, Tóth R, Nagy F et al. (2005) Plant circadian
clocks increase photosynthesis, growth, survival, and competitive advantage. Science, 309, 630–633.
Elliott S, Baker PJ, Borchert R (2006) Leaf flushing during the dry season: the paradox
of Asian monsoon forests. Global Ecology and Biogeography, 15, 248–257.
Gotelli NJ, Graves GR (1996). The temporal niche. In: Null Models in Ecology (eds Gotelli NJ, Graves GR), pp. 95–111. Smithsonian Institution, Washington, DC.
Hoffman AA, Sgrò CM (2011) Climate change and evolutionary adaptation. Nature,
470, 479–485.
Huete AR, Didan K, Shimabukuro YE, Ratana P, Saleska SR, Hutyra LR et al. (2006)
Amazon rainforests green-up with sunlight in dry season. Geophysical Research Letters, 33, 4 pp.
Inouye DW (2008) Effects of climate change on phenology, frost damage, and floral
abundance of montane wildflowers. Ecology, 89, 353–362.
Jackson ST, Betancourt JL, Booth RK, Gray ST (2009) Ecology and the ratchet of
events: climate variability, niche dimensions, and species distributions. Proceedings
of the National Academy of Sciences, 106, 19685–19692.
Jansson R, Dynesius M (2002) The fate of clades in a world of recurrent climatic
change: Milankovitch oscillations and evolution. Annual Review of Ecology and Systematics, 33, 741–777.
Jentsch A, Kreyling J, Boettcher-Treschkow J, Beierkuhnlein C (2009) Beyond gradual
warming: extreme weather events alter flower phenology of European grassland
and heath species. Global Change Biology, 15, 837–849.
Johnson SD (1993) Climatic and phylogenetic determinants of flowering seasonality
in the Cape Flora. Journal of Ecology, 81, 567–572.
Kochmer JP, Handel SN (1986) Constraints and competition in the evolution of flowering phenology. Ecological Monographs, 56, 303–325.
Kraft NJB, Cornwell WK, Webb CO, Ackerly DD (2007) Trait evolution, community
assembly, and the phylogenetic structure of ecological communities. American Naturalist, 170, 271–283.
Lack AJ (1982) Competition for pollinators in the ecology of Centaurea scabiosa L. and
Centaurea nigra L. I. variation in flowering time. New Phytologist, 91, 297–308.
Larcher W (2003). Plants under stress. In: Physiological Plant Ecology: Ecophysiology and
Stress Physiology of Functional Groups, 4th ed. (ed. Larcher W), pp 345–450. Springer,
Berlin.
© 2011 Blackwell Publishing Ltd, Global Change Biology, 17, 3633–3643
P R E D I C T I N G P H E N O L O G Y 3643
Levin DA (2006) Flowering phenology in relation to adaptive radiation. Systematic
Botany, 31, 239–246.
Schwartz MD, Reiter BE (2000) Changes in North American spring. International Journal of Climatology, 20, 929–932.
Lichstein JW, Dushoff J, Ogle K, Chen A, Purves ADW, Caspersen JP, Pacala SW
(2010) Unlocking the forest inventory data: relating individual-tree performance to
unmeasured environmental factors. Ecological Applications, 20, 684–699.
Losos JB (2008) Phylogenetic niche conservatism, phylogenetic signal and the relationship between phylogenetic relatedness and ecological similarity among species. Ecology Letters, 11, 995–1003.
Martin PR, Bonier F, Moore IT, Tewksbury JT (2009) Latitudinal variation in the asyn-
Schwartz MD, Ahas R, Aasa A (2006) Onset of spring starting earlier across the
Northern Hemisphere. Global Change Biology, 12, 343–351.
Stenseth NC, Mysterud A (2005) Weather packages: finding the right scale and composition of climate in ecology. Journal of Animal Ecology, 74, 1195–1198.
Suding KN, Lavorel S, Chapin FS, Cornelissen JHC, Dı́az S, Garnier E et al. (2008)
Scaling environmental change through the community-level: a trait-based
response-and-effect framework for plants. Global Change Biology, 14, 1125–1140.
chrony of seasons: implications for higher rates of population differentiation and
speciation in the tropics. Ideas in Ecology and Evolution, 2, 9–17.
Mazer SJ (1990) Seed mass of Indiana Dune genera and families: taxonomic and ecological correlates. Evolutionary Ecology, 4, 326–357.
McGill BJ, Enquist BJ, Weiher E, Westoby M (2006) Rebuilding community ecology
from functional traits. Trends in Ecology & Evolution, 21, 178–185.
Thackeray SJ, Sparks TH, Frederiksen M, Burthe S, Bacon PJ, Bell JR et al. (2010) Trophic level asynchrony in rates of phenological change for marine, freshwater and
terrestrial environments. Global Change Biology, 16, 3304–3313.
van der Valk AG (1981) Succession in wetlands – a Gleasonian approach. Ecology, 62,
688–696.
Venable DL (2007) Bet hedging in a guild of desert annuals. Ecology, 88, 1086–1090.
Menzel A, Sparks TH, Estrella N, Koch E, Aasa A, Ahas R et al. (2006) European phenological response to climate change matches the warming pattern. Global Change
Biology, 12, 1969–1976.
Miller-Rushing AJ, Høye TT, Inouye DW, Post E (2010) The effects of phenological
mismatches on demography. Philosophical Transactions of the Royal Society B: Biological Sciences, 365, 3177–3186.
Mittelbach GG, Schemske DW, Cornell HV, Allen AP, Brown JM, Bush MB et al.
Vicente-Serrano SM, Delbart N, Toan TL, Grippa M (2006) El Niño–Southern Oscillation influences on the interannual variability of leaf appearance dates in central
Siberia. Geophysical Research Letters, 33, 4 pp.
Visser ME (2008) Keeping up with a warming world; assessing the rate of adaptation
to climate change. Proceedings of the Royal Society B: Biological Sciences, 275, 649–659.
Vose RS, Easterling DR, Gleason B (2005) Maximum and minimum temperature
trends for the globe: an update through 2004. Geophysical Research Letters, 32, 5 pp.
(2007) Evolution and the latitudinal diversity gradient: speciation, extinction and
biogeography. Ecology Letters, 10, 315–331.
Morales MA, Dodge GJ, Inouye DW (2005) A phenological mid-domain effect in
flowering diversity. Oecologia, 142, 83–89.
Morin X, Viner D, Chuine I (2008) Tree species range shifts at a continental scale: new
predictive insights from a process-based model. Journal of Ecology, 96, 784–794.
Webb CO (2000) Exploring the phylogenetic structure of ecological communities: an
example for rain forest trees. American Naturalist, 156, 145–155.
Webb CO, Ackerly DD, McPeek MA, Donoghue MJ (2002) Phylogenies and community ecology. Annual Review of Ecology and Systematics, 33, 475–505.
Wiens JJ, Ackerly DD, Allen AP, Anacker BL, Buckley LB, Cornell HV et al. (2010)
Niche conservatism as an emerging principle in ecology and conservation biology.
Morin X, Lechowicz MJ, Augspurger C, O’Keefe J, Viner D, Chuine I (2009) Leaf phenology in 22 North American tree species during the 21st century. Global Change
Biology, 15, 961–975.
Mouradov A, Cremer F, Coupland G (2002) Control of flowering time: interacting
pathways as a basis for diversity. Plant Cell, 14, S111–S130.
Oberrath R, Böhning-Gaese K (2002) Phenological adaptation of ant-dispersed plants
to seasonal variation in ant acitivty. Ecology, 83, 1412–1420.
Ecology Letters, 13, 1310–1324.
Wilczek AM, Roe JL, Knapp MC, Cooper MD, Lopez-Gallego C, Martin LJ et al.
(2009) Effects of genetic perturbation on seasonal life history plasticity. Science,
323, 930–934.
Williams LJ, Bunyavejchewin S, Baker PJ (2008) Deciduousness in a seasonal tropical
forest in western Thailand: interannual and intraspecific variation in timing, duration and environmental cues. Oecologia, 155, 571–582.
Ollerton J, Lack AJ (1992) Flowering phenology: an example of relaxation of natural
selection? Trends in Ecology & Evolution, 7, 274–276.
Olsson K, Agren J (2002) Latitudinal population differentiation in phenology, life history and flower morphology in the perennial herb Lythrum salicaria. Journal of Evolutionary Biology, 15, 983–996.
Parmesan C (2007) Influences of species, latitudes and methodologies on estimates of
Willis CG, Ruhfel B, Primack RB, Miller-Rushing AJ, Davis CC (2008) Phylogenetic
patterns of species loss in Thoreau’s woods are driven by climate change. Proceedings of the National Academy of Sciences USA, 105, 17029–17033.
Wright SJ, Calderon O (1995) Phylogenetic patterns among tropical flowering phenologies. Journal of Ecology, 83, 937–948.
Yang LH, Rudolf VHW (2010) Phenology, ontogeny and the effects of climate change
phenological response to global warming. Global Change Biology, 13, 1860–1872.
Peñuelas J, Rutishauser T, Filella I (2009) Phenology feedbacks on climate change. Science, 324, 887–888.
Peterson TC, Vose RS (1997) An overview of the Global Historical Climatology Network temperature database. Bulletin of the American Meteorological Society, 78, 2837–
2849.
Post E, Stenseth NC (1999) Climatic variability, plant phenology, and northern ungulates. Ecology, 80, 1322–1339.
Rabinowitz D, Rapp JK, Sork VL, Rathcke BJ, Reese GA, Weaver JC (1981) Phenological properties of wind- and insect-pollinated prairie plants. Ecology, 62, 49–56.
Reich P, Borchert R (1984) Water stress and tree phenology in a tropical dry forest in
the lowlands of Costa Rica. Journal of Ecology, 72, 61–74.
Revell LJ, Harmon LJ, Collar DC (2008) Phylogenetic signal, evolutionary process,
and rate. Systematic Biology, 57, 591–601.
Richardson AD, Hollinger DY, Dail DB, Lee JT, Munger JW, O’keefe J (2009) Influence
of spring phenology on seasonal and annual carbon balance in two contrasting
New England forests. Tree Physiology, 29, 321–331.
Sakai S (2001) Phenological diversity in tropical forests. Population Ecology, 43, 77–86.
Sargent RD, Ackerly DD (2008) Plant-pollinator interactions and the assembly of
plant communities. Trends in Ecology & Evolution, 23, 123–130.
van Schaik CP, Terborgh JW, Wright SJ (1993) The phenology of tropical forests:
adaptive significance and consequences for primary consumers. Annual Review of
Ecology and Systematics, 24, 353–377.
© 2011 Blackwell Publishing Ltd, Global Change Biology, 17, 3633–3643
on the timing of species interactions. Ecology Letters, 13, 1–10.
Supporting Information
Additional Supporting Information may be found in the
online version of this article:
Appendix S1. Literature review description and references.
Appendix S2. References and methods for Figs 2 and 5.
Appendix S3. Author contributions.
Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied
by the authors. Any queries (other than missing material)
should be directed to the corresponding author for the
article.